I had made a half-wave rectifier circuit in LTSpice, consisting of a sinusoidal voltage input, a diode that would be forward-biased during the positive half-cycles of the input, a 1k ohm resistor and a ground node. After wiring, I ran this circuit through a DC Sweep analysis, varying the input voltage linearly with a Start value of -400 V, a Stop value of 400 V and an Increment of 100 V (this was done to obtain the voltage transfer characteristics of the circuit, if there is a better way of obtaining these characteristics, please let me know). The plot was the voltage across the resistor vs. the varying input. The plot was as expected assuming an ideal diode model.
Then, I wanted to observe the transfer characteristics of the same circuit with the diode forward-biased during negative half-cycles. Now when I tried to obtain the same plot (using the same DC Sweep analysis), it came out strange:

(SCHEMATIC)
​

UNEXPECTED PLOT



​

The plot is strange around the -80 V - +80 V region of the horizontal axis. The trace is supposed to remain flat till 0 V and then, with a slope of -1, enter the 3rd quadrant as a straight line.

I tried again with a simple DC source, a source with no specifications, different Start/Stop/Increment values, a "signal" source, a source in series with a resistor (to check whether the source was at fault), etc. (all simulated using DC Sweep analysis). This leads me to conclude that most probably the DC Sweep analysis is not appropriate for plotting transfer characteristics or I am not defining the source properly.

Please explain why this is happening and how to obtain the expected transfer characteristics using LTSpice. Thank you.

 

Try turning the diode around.

 

Change the increment value to 10mV.
Why did you use such a large increment value?
That value determines the precision of your measurement.

 

Try turning the diode around.

The circuit is supposed to allow only -ve half cycles of input voltage to appear across the resistor. If I turn the diode around, it blocks all -ve half cycles and allows only +ve half cycles. This circuit works fine, but it is not what I want. Thank you.

 

Change the increment value to 10mV.
Why did you use such a large increment value?
That value determines the precision of your measurement.

Thank you for your response. After I changed the interval to 10mV, the plot is as expected. Could you kindly explain how DC Sweep analysis basically works and why a change in the "interval" parameter resulted in two totally different plots? Shouldn't the plots be the same regardless of the "interval"? Also, I used the same "interval" value of 100 V for analyzing the +ve half-cycle rectifier and the plot came out as expected, why this change when I reverse the diode?

 

This circuit works fine, but it is not what I want. Thank you.

My bad . I wasn't paying close enough attention...

 

My bad . I wasn't paying close enough attention...

Thank you for your response. "crutschow" 's answer is giving the expected plot. But I don't know why reducing the "interval" results in the expected plot.

 

Thank you for your response. After I changed the interval to 10mV, the plot is as expected. Could you kindly explain how DC Sweep analysis basically works and why a change in the "interval" parameter resulted in two totally different plots? Shouldn't the plots be the same regardless of the "interval"? Also, I used the same "interval" value of 100 V for analyzing the +ve half-cycle rectifier and the plot came out as expected, why this change when I reverse the diode?

The "interval" is the values it uses to do the measurements.
Obviously if it uses a 100V interval than there are no measurements in between and so you know nothing about what happens in that interval. So for the plot it just draws a straight line between those two points.

Think if you were doing this measurement by hand and changing the power supply by 100V between measurements. Not much information there with just 9 measurements.

In my simulation I got the same glitch for either orientation of the diode for a 100V increment.

 

Thank you

The "interval" is the values it uses to do the measurements.
Obviously if it uses a 100V interval than there are no measurements in between and so you know nothing about what happens in that interval. So for the plot it just draws a straight line between those two points.

Think if you were doing this measurement by hand and changing the power supply by 100V between measurements. Not much information there with just 9 measurements.

In my simulation I got the same glitch for either orientation of the diode for a 100V increment.

Thank you very much for your answer.